LOGIN
Bi Sh
Design Invention Research
 · KathmanduNepal
Edit
Delete
Share
Report
IDEA

Idea
#AI, #MI

Cognitive Map
Emotional Map
Dynamic Genetic Map
Engram (ActiveMemory) Map

Cognitive mapping of human brain is already a success decades back! May b we might make it more efficient,
but as pointed by various critiques… cognitive mapping is just a past of complete picture…
Emotional, Dynamic Genetic, Memory, Active Engram maps might help it b more like humans,
But still further, experts suggests that Machine Intelligence is very different than AI that tries to b like Humans, which if done properly will surpass Human Intelligence in various areas.
So, in conclusion, if we think from the perspective of making the machine like humans… our end result will b like humans… but if we choose Machine Intelligence, that r not strictly bounded to human intelligence… will go beyond the target!

1 like 
Like
Share
Rob Freeman
Label resister (Project Leader)
 · Los AngelesU.S.
Edit
Delete
Share
Report

My project has much more specific goals than this Bi Sh. All I want to do is find structure in a sequence.

It is amusing to speculate more broadly on intelligence. I’ve speculated myself that there might be non-human “intelligences”. But I was able to be clear about what I meant by that, in terms of searches for different kinds of structure.

I think it is only useful to talk about the problem in terms of searches for structure at this point.

Anyway, mine is a clearly defined project, based on the idea that finding structure in sequences is fundamental. It is clearly differentiated from other such projects. Most of which can also been seen as searches for structure. The point of differentiation is limited to ideas about how we should parametrize structure, and most specifically in my case, ideas about the complexity of structure which is possible.

I think it is true to say all, 99% anyway, of contemporary AI projects talk about finding structure.

So discussion which does not phrase the problem in those terms, I don’t think is useful. Things become too general, and you start philosophizing instead of building.

We already have a model. We are testing it, human or non-human.

But I’m curious what you mean by “cognitive mapping”. Do you have a reference for the work “decades back” so I can see exactly what you mean?

1 like 
Bi Sh
Design Invention Research
 · KathmanduNepal
Edit
Delete
Share
Report

Rob Freeman…
i replied this days back… but it popped up now… i dont this this is relevant but its posting…
yeah, i know im not being specific n not focusing on the task at hand… just seeing it from outset… will try to delve deeper soon…
Decades back i happen to see a video called “Blue Brain Project” which claimed that for the first time they r able to locate the brain map of experience… neural correlates of experience… for example observing a beautiful rose with smell… the project claimed to locate all the neural correlates… cognitive map is not the term i know… but amazingly the whole thing got vanished from youtube never to found again! i’vent searched it for long… will definitely post it if i get some link to it. as far as i remember it was a huge project with lot of scientists n institutes, may b European one i guess. I can only talk broadly about Structures in AI… suppose the implementation of Structure generates the desired intelligent behavior… so the real situation is Dynamic… not some static blueprint. But the Static (Structure, if im getting it right) leads to Dynamic Functioning. Evolutionary AI algorithms (structures) can simulate Evolution but can not predict whats coming! I know im again dwelling in philosophical realms… will definitely manage more time to get everything updated.

0 like 
Rob Freeman
Label resister (Project Leader)
 · Los AngelesU.S.
Edit
Delete
Share
Report

OK, thanks. Blue Brain Project was your source.

I don’t think any maps are a solved problem at all. The best general reference I know currently might be this work:

Tom Mitchell, Carnegie Mellon University
https://youtu.be/8zcBr6bFk1A?t=37m

The best they have for new structures (his “generative” theory) are some naive mixtures.

Here’s some notes I made on this talk elsewhere. For your general interest:

'Going back before minute 37, his “generative” theory is interesting, around minute 24. It looks very like the first, linear, approximation of context generalization for words in Yoshua Bengio’s first “neural language models.”

Nice that he too is using context generalization to mix his representations for his first attempt at a generative model. Context generalization rules.

But what I would highlight from this presentation is in the study of representation combination, around minute 53:32.

It is observed that the representation of an adjective changes when a noun is presented. It can be decoded as an adjective, but it is a different representation.’

1 like 
Bi Sh
Design Invention Research
 · KathmanduNepal
Edit
Delete
Share
Report

my previous comment also got lost…
Regarding Text (Language) or “Generative”, it might improve if Chomsky’s work in this field is applied. I think u know it already… few essence i get when i reflect back are…
a. Learned Imprints in brain
b. Innate Imprints in brain, n as its innate the “Innate Engram” in brain will definitely have its correlation in Gene too.

0 like 
Bi Sh
Design Invention Research
 · KathmanduNepal
Edit
Delete
Share
Report

few attempts,

  1. Neural correlate of Beauty:http://journals.plos.org/plosone/article?id=10.1371/journal.pone.0031248
  2. Neural Anesthetics http://jn.physiology.org/content/jn/91/4/1699.full.pdf
  3. Blue Brain Project is now Collaborationg Globally for Human Brain Project… n as this latest video says :https://www.youtube.com/watch?v=37SFQri8g_U
    its far from over…
    …very strange, because i remember seeing the Graphic Visualization of exact neural correlate of experience by then itself!
0 like 
Leave a reply...
DISCOVER
CHAT
ALERTS
-2
DISCUSSIONS
-1
FEED